JP2008282097A - Collision risk degree estimating apparatus and driver supporting apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a collision risk degree estimating apparatus capable of estimating a collision risk degree with high reliability, and to provide a driver supporting apparatus capable of supporting an operation of a driver with high reliability so as to avoid collision. <P>SOLUTION: A pedestrian detecting unit 42 detects a position of a pedestrian who moves in periphery of a cruising vehicle, a moving speed of the pedestrian, and a moving direction of the pedestrian, and a barrier detecting unit 46 detects a position of a barrier which prevents movement of the pedestrian. A crossing intent estimating unit 44 and a collision risk degree estimating unit 48 estimate collision risk degree about collision between the pedestrian and the vehicle based on the detected position, moving speed, and moving direction of the pedestrian, and a detected position of the barrier. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a collision risk estimation device and a driver assistance device, and more particularly to a collision risk estimation device that estimates a collision risk that a pedestrian and a vehicle collide, and a driver assistance device that includes the collision risk estimation device. About.

  Conventionally, various techniques for preventing a collision accident between a vehicle and a pedestrian entering the front of the vehicle have been proposed.

  For example, in Patent Document 1, a lateral moving object on a traveling path of the host vehicle is detected, a lateral moving speed, a position, and a size of the detected lateral moving object are acquired, and the acquired lateral moving speed, position, and size are acquired. Based on the above, a technique for determining a risk level related to traveling of the host vehicle is disclosed.

  In addition, in the patent document 2, the present applicant recognizes an object in the environment around the vehicle using a CCD camera or a laser transmitter / receiver, and further determines the direction with respect to the traveling direction of the vehicle for each type of the recognized object. Technology that recognizes the speed, size, and position, calculates the risk according to the type of object using the corresponding risk parameters, and outputs the calculated risk by image or sound Disclosed.

Further, Patent Document 3 detects a pedestrian around the vehicle and calculates a predicted course of the detected pedestrian while detecting a moving body around the vehicle and calculates a predicted course of the detected moving body. Based on the calculated predicted course of the pedestrian and the predicted path of the moving object, the possibility of collision between the pedestrian and the moving object is calculated. A technique for notifying a body of a collision avoidance message is disclosed.
JP 2000-251200 A JP 2003-81039 A JP 2006-31443 A

  However, when the pedestrian around the vehicle is detected using the techniques of Patent Documents 1 to 3 and the collision risk of collision between the vehicle and the pedestrian is estimated, for example, the pedestrian walks. Even if there is a barrier between the sidewalk and the roadway (roadside) that prevents the movement of pedestrians such as guard rails, the collision between the vehicle and the pedestrian is low. There is a problem that the risk is estimated to be high and the reliability of the estimated collision risk is low.

  The present invention has been made in order to solve the above-described problems, and a collision risk estimation device capable of estimating a collision risk with high reliability, and driving of a driver so as to avoid a collision with high reliability. An object of the present invention is to provide a driver support apparatus that can support the above.

  In order to achieve the above object, the invention described in claim 1 includes pedestrian detection means for detecting a position of a pedestrian moving around a traveling vehicle, a moving speed of the pedestrian, and a moving direction of the pedestrian. A barrier detecting means for detecting a position of the barrier that prevents the movement of the pedestrian, a position, a moving speed, and a moving direction of the pedestrian detected by the pedestrian detecting means, and the barrier for the pedestrian and the vehicle. Estimation means for estimating a collision risk that the pedestrian and the vehicle collide based on the position of the barrier detected by the detection means.

  In the first aspect of the invention, the position of the pedestrian moving around the traveling vehicle, the moving speed of the pedestrian, and the moving direction of the pedestrian are detected by the pedestrian detection means, and the barrier detection means is used. The position of the barrier that prevents the movement of the pedestrian is detected.

  In the present invention, based on the position, the moving speed, and the moving direction of the pedestrian detected by the pedestrian detecting means and the position of the barrier detected by the barrier detecting means for the pedestrian and the vehicle by the estimating means. Thus, the risk of collision between the pedestrian and the vehicle is estimated.

  Thus, according to the first aspect of the invention, the position of the pedestrian moving around the traveling vehicle, the moving speed of the pedestrian and the moving direction of the pedestrian are detected, and the movement of the pedestrian is detected. Collision risk that pedestrian and vehicle collide based on detected pedestrian position, moving speed, and moving direction and detected barrier position with respect to pedestrian and vehicle Since the degree is estimated, the collision risk can be estimated with high reliability.

  According to the first aspect of the present invention, as in the second aspect of the present invention, the surrounding situation of the vehicle is obtained by performing at least one of imaging of the periphery of the vehicle, scanning with a laser beam, and scanning with a radio wave. An acquisition means for acquiring the situation information shown, wherein the pedestrian detection means determines the position of the pedestrian, the movement speed of the pedestrian, and the pedestrian from the surrounding situation indicated by the situation information acquired by the acquisition means. And the barrier detecting means may detect the position of the barrier from the surrounding situation indicated by the situation information acquired by the acquiring means.

  The barrier according to the present invention is, as in the third aspect of the invention, at least one of a guard fence, a structure that prevents the pedestrian including a median, a traffic condition of an oncoming lane, and a bad road surface. There may be.

  The estimating means according to the present invention may be configured such that, when the barrier detected by the barrier detecting means is located between the pedestrian and the vehicle, as in the invention according to claim 4, the barrier is the pedestrian. It is preferable that the collision risk is estimated to be lower than that between the vehicle and the vehicle.

  Further, according to the estimation means of the present invention, the movement direction of the pedestrian detected by the pedestrian detection means is substantially parallel to the movement direction of the vehicle, and the barrier detection is performed. When the barrier detected by the means is located in front of the pedestrian in the moving direction, it is preferable to estimate the collision risk higher than when the barrier is not positioned in front of the moving direction.

  In the present invention, as in the invention described in claim 6, the pedestrian detection means further detects the body direction or the face direction of the pedestrian, and the estimation means is detected by the pedestrian detection means. When the detected body orientation or face orientation of the pedestrian is front or back relative to the vehicle, the risk of collision is lower than when the orientation is other than front or back relative to the vehicle. It is preferable to estimate.

  Further, as in the invention according to claim 7, the present invention further comprises position information acquisition for acquiring position information indicating a position of a predetermined use target that can be used by a pedestrian, and the estimation means includes When the position of the use target indicated by the position information acquired by the position information acquisition is located on the opposite side across the road on which the vehicle travels with respect to the position of the pedestrian, the position of the use target is It is preferable that the collision risk is estimated to be higher than when it is not located on the opposite side. The acquisition of the position information by the acquisition of the position information includes acquisition by reading the position information from a storage unit that stores the position information in advance, and acquisition by receiving the position information from the outside. In addition, the above objects of use include various amusement facilities such as a bus stop, a station, a school, a store, a company, a store, a movie theater, and a zoo.

  Moreover, the said estimation means of Claim 7 is the time slot | zone used frequently when the said utilization object is the time slot | zone used frequently by the said pedestrian like invention of Claim 8. It is preferable to estimate the collision risk higher than in other cases.

  Further, as in the invention according to claim 9, the estimating means of the present invention is a crossing in which the pedestrian and the vehicle collide with each other when the pedestrian moves in a crossing direction with respect to a road on which the vehicle travels. Calculating the direction collision probability, calculating the translation direction collision probability that the pedestrian moves in the translation direction with respect to the road and the pedestrian collides with the vehicle, and calculates the cross direction collision probability and the translation direction. The collision risk may be estimated by adding the collision probabilities.

  Furthermore, the pedestrian detection means of the present invention, as in the invention of claim 10, excludes a region where movement of the pedestrian is hindered by the barrier detected by the barrier detection means from being a pedestrian detection target. Also good.

  On the other hand, the invention according to claim 11 is such that the collision risk estimated by the collision risk estimation device according to any one of claims 1 to 10 and the collision risk estimation device is not more than a predetermined value. Deriving means for deriving the speed of the vehicle as an appropriate speed, and display means for displaying the appropriate speed derived by the deriving means.

  Therefore, according to the invention described in claim 11, the collision risk is estimated with high reliability by the collision risk estimation device, and the collision risk estimated by the collision risk estimation device by the derivation means is less than a predetermined value. Since the vehicle speed is derived as an appropriate speed and the appropriate speed derived by the deriving means is displayed by the display means, the driving of the driver can be assisted to avoid a collision with high reliability. . The display by the display means includes visible display by a display device or the like, audible display by a speaker or the like.

  On the other hand, in the invention described in claim 12, the collision risk estimation apparatus according to any one of claims 1 to 10 and the collision risk estimated by the collision risk estimation apparatus are determined in advance. Warning means for warning the driver of the vehicle when the value is larger than

  Therefore, according to the twelfth aspect of the present invention, the collision risk estimation device estimates the collision risk with high reliability, and the warning means predetermines the collision risk estimated by the collision risk estimation device. When the value is larger than the above value, a warning is given to the driver of the vehicle, so that it is possible to assist the driving of the driver so as to avoid the collision with high reliability.

  On the other hand, in the invention described in claim 13, the collision risk estimation device according to any one of claims 1 to 10 and the collision risk estimated by the collision risk estimation device are determined in advance. Braking amount control means for controlling to increase the braking amount of the brake with respect to the brake operation to the brake pedal provided in the vehicle.

  Therefore, according to the invention described in claim 13, the collision risk estimation device estimates the collision risk with high reliability, and the braking risk control device estimates the collision risk estimated by the collision risk estimation device in advance. When the value is larger than the set value, the brake is controlled so as to increase the braking amount for the brake operation to the brake pedal provided in the vehicle, so that the driver is driven to avoid the collision with high reliability. Can help.

  As described above, according to the present invention, the position of the pedestrian moving around the traveling vehicle, the moving speed of the pedestrian, and the moving direction of the pedestrian are detected, and the barrier that prevents the pedestrian from moving. The collision risk of collision between the pedestrian and the vehicle is determined based on the detected position, moving speed, and moving direction of the pedestrian and the detected position of the barrier with respect to the pedestrian and the vehicle. Since the estimation is performed, the collision risk can be estimated with high reliability.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the following description, the present invention is applied to a driver assistance device that is mounted on a vehicle, estimates a collision risk that a pedestrian around the vehicle collides with the vehicle, and issues a warning according to the estimated collision risk. This will be described as an example.

FIG. 1 shows a schematic configuration of a driver support apparatus 10 according to the present embodiment.
As shown in the figure, the driver assistance apparatus 10 includes a laser radar 12 that acquires situation information indicating the surrounding situation of the vehicle by scanning the surroundings of the vehicle with laser light, a display 14 that displays various information, and a laser. Collision in which the pedestrian and the position of the barrier that prevents the movement of the pedestrian are detected based on the situation information acquired by the radar 12, and the pedestrian and the vehicle collide based on the positional relationship between the detected pedestrian and the barrier. And an apparatus main body 20 that estimates a risk and issues a warning according to the estimated collision risk.

  The laser radar 12 is attached to the front side of the vehicle such as a front grill or a bumper of the vehicle. The laser radar 12 measures the time from when the pulse laser beam is emitted until it is reflected back to the three-dimensional object, and obtains the distance from the three-dimensional object by calculating the product of the measured time and the speed of light. The shape of the three-dimensional object in front of the vehicle is measured by scanning the laser beam in the horizontal direction with a scanning mechanism.

  As shown in FIG. 1, the apparatus main body 20 includes a CPU (Central Processing Unit) 22 that controls the operation of the entire driver support apparatus 10, and a RAM (Random Access Memory) used as a work area when the CPU 22 executes various processing programs. ) 24, a ROM (Read Only Memory) 26 in which various control programs including various collision risk estimation processing programs, which will be described later, various parameters, and the like are stored in advance, and an HDD (Hard Disk Drive) 28 in which various information is stored. A radar control unit 30 for controlling the scanning operation of the laser radar 12, a display control unit 32 for controlling display of various information such as a screen and a message on the display 14, a speaker 34 for giving a warning to the driver, It has.

  The CPU 22, RAM 24, ROM 26, HDD 28, radar control unit 30, display control unit 32, and speaker 34 are connected to each other via a system bus BUS.

  Therefore, the CPU 22 accesses the RAM 24, the ROM 26, and the HDD 28, controls the scanning operation by the laser radar 12 through the radar control unit 30, and controls the display of various information on the display 14 through the display control unit 32. Control of output of sound from the speaker 34 can be performed.

  As shown in FIG. 2A, the radar control unit 30 controls the scanning operation performed by the laser radar 12 so that the polar coordinate expression represented by the scan angle and distance with the vehicle position as the origin is shown in FIG. The situation information indicating the surrounding situation of the vehicle is acquired. As shown in FIG. 2B, the radar control unit 30 converts the acquired polar coordinate expression situation information into situation information of an orthogonal coordinate system in which the road is looked down from above, and outputs it.

  FIG. 3 is a functional block diagram showing a functional configuration of the driver support apparatus 10 according to the present embodiment.

  As shown in the figure, the driver assistance device 10 includes a grouping processing unit 40, a pedestrian detection unit 42, a crossing intention estimation unit 44, a barrier detection unit 46, and a collision risk estimation unit 48. ing.

  The grouping processing unit 40 is connected to the radar control unit 30, the pedestrian detection unit 42, and the barrier detection unit 46, and situation information is input from the radar control unit 30.

Here, the following objects are mainly present around the traveling vehicle.
○ Moving objects ・ Vehicles (leading vehicle, oncoming vehicle, motorcycle)
・ Pedestrian ○ Stationary objects ・ Parking vehicles ・ Roadside objects (electric poles, protective fences, etc.) existing between the roadway and the sidewalk
・ Block structures (buildings, houses, walls, etc.)

  When the position of the solid object indicated by the situation information is plotted, the portion where the solid object is spatially continuous can be regarded as any of the above objects, and the object existing on the road side is directed to the pedestrian's roadway. Candidate barriers such as protective fences that prevent the entry of

  The grouping processing unit 40 plots the positions of the three-dimensional objects indicated by the situation information, and groups the regions where the three-dimensional objects are continuous as objects. The laser radar 12 may lose some data. For this reason, for example, when the distance between adjacent three-dimensional objects is 30 cm or less in the orthogonal coordinate system with reference to the thickness of the utility pole of about 30 cm, they are combined and regarded as the same object.

  The pedestrian detection unit 42 tracks an object having a predetermined size (in this case, about 0.4 m in width) about the width of the pedestrian from each object grouped by the grouping processing unit 40 in the time direction as a pedestrian candidate. A person moving on a road at a predetermined speed (for example, about 1.2 [m / s]) at which the pedestrian moves is identified as a pedestrian. And the pedestrian detection part 42 detects the position of the specified pedestrian with respect to the own vehicle, the moving speed of the said pedestrian, and the moving direction of the said pedestrian. It should be noted that the image processing is limited to a region where an object having a width of about 0.4 m is present, and the object is stopped using a known pattern identification method (SVM (Support Vector Machine) or cascade type identification). The pedestrian who is present may be further specified.

  Moreover, the pedestrian detection part 42 detects a pedestrian's attitude | position (body direction with respect to the own lane) by further implementing pattern identification with respect to the area | region identified as the pedestrian.

  As this pattern identification method, for example, pattern data for each posture created by scanning a pedestrian whose pedestrian is facing back and front and a pedestrian facing sideways with the laser radar 12 when viewed from the own vehicle is used. It is stored in the HDD 28 in advance, the area identified as a pedestrian and each pattern data are compared to derive the similarity, and the posture corresponding to the pattern data having the highest similarity is detected as the pedestrian's posture. do it.

  The HDD 28 stores in advance map information including position information indicating positions of predetermined use targets that can be used by pedestrians such as bus stops, stations, schools, shops, companies, and pedestrian crossings. In the present embodiment, the map information is stored in advance in the HDD 28. However, for example, when the driver support apparatus 10 is linked with the car navigation system, the map information held by the car navigation system is read out. It may be used. Further, for example, signal state information indicating surrounding map information and signal state may be obtained by road-to-vehicle communication.

  The crossing intention estimation unit 44 is a position of a pedestrian detected by the pedestrian detection unit 42, the posture of the pedestrian and the position of a bus stop, a station, a school, a store, or a company in the vicinity indicated by the map information stored in the HDD 28. The crossing intention of the pedestrian is estimated from the positional relationship with

  That is, when the direction of the pedestrian is sideways with respect to the vehicle, there is a high possibility that the pedestrian jumps out on the road because the pedestrian intends to cross. Further, when the pedestrian is facing the front of the vehicle, the pedestrian recognizes the presence of the vehicle, and therefore, the possibility that the pedestrian jumps out on the road is low. Furthermore, when the pedestrian is facing the rear of the vehicle, the pedestrian may not jump to the road because the pedestrian does not recognize the presence of the vehicle but moves in a direction parallel to the vehicle direction. Low.

  In addition, for example, a bus stop, a station, a school, or a company is located on the opposite side of the road on which the vehicle runs with respect to the position of the pedestrian, and commuting / commuting hours (for example, 8:00 to 9:00) In this case, since the pedestrian goes to a bus stop, a station, a school, or a company, the pedestrian is likely to jump out on the road. In addition, for example, if a store is located on the opposite side of the pedestrian across the road and is in a shopping time zone (for example, from 15:00 to 18:00), the pedestrian goes to the store. Therefore, there is a high possibility that pedestrians jump out on the road. Furthermore, if the distance between the nearest pedestrian crossing and the pedestrian indicated by the map information stored in the HDD 28 is not less than a set value (for example, 30 m or more), the pedestrian is likely to jump out of the pedestrian crossing. .

  Further, in the time zone of midnight (for example, 23:00 to 3) or early morning (for example, 3 to 6), there is a high possibility that a pedestrian jumps out on the road because the number of vehicles traveling on the road is small.

  In the driver assistance device 10 according to the present embodiment, in order to estimate the pedestrian's crossing intention as described above, the position of the pedestrian with respect to the vehicle, the position of the surrounding bus stop, station, school, store, company For each item such as a relationship and a time zone, information indicating a contribution according to the item is stored in the HDD 28 in advance. The degree of contribution for each item may be set based on, for example, the result of observing the behavior of a pedestrian in an actual road environment, or may be set based on the knowledge of a good driver.

  The crossing intention estimation unit 44 obtains the sum of the contributions of the items corresponding to the situation detected by the pedestrian detection unit 42 from the information indicating the contributions of each item stored in the HDD 28, so that the pedestrian crossing intention Is estimated.

  The pedestrian detection unit 42 detects the size (height) of the area identified as a pedestrian and the pedestrian's stride, and the crossing intention estimation unit 44 detects the pedestrian's size and stride detected by the pedestrian detection unit 42. The age of the pedestrian is estimated from the above, and when it is estimated that the child is old or old, the pedestrian crossing intention may be estimated by determining that the probability of popping out is high without confirming the approach.

  Further, although the pedestrian detection unit 42 according to the present embodiment has been described with respect to a case where a barrier is detected using the laser radar 12, for example, two units are used together with the laser radar 12 or instead of the laser radar 12. The three-dimensional object may be detected by continuously capturing images with the cameras and performing image processing by stereo vision on the images obtained by the imaging by the cameras. In this case, the height information of the three-dimensional object can be obtained by grouping the three-dimensional object and statistically obtaining the distribution in the height direction. In this case, in Japan, the height of the protective fence provided on the road is standardized to about 80 cm. Therefore, the protective fence can be determined based on the height information. In addition, the pedestrian detection unit 42 detects a face pattern from the upper part of the pedestrian detection area of the image obtained by imaging, and determines that the face direction is detected when the face pattern is detected, When the face pattern is not detected, it is determined that the direction of the host vehicle is not seen, thereby detecting the direction of the face of the pedestrian with respect to the host vehicle, and the crossing intention estimation unit 44 does not see the host vehicle. If it is determined that the pedestrian crossing intention is estimated, it is determined that the possibility of jumping out is high, and that it is determined that the possibility of jumping out is low when it is determined that the host vehicle is being viewed.

  The barrier detection unit 46 detects, as a barrier, an object that hinders the movement of the pedestrian from each object grouped by the grouping processing unit 40. This barrier prevents the movement of pedestrians such as structures on the road (guard fences, median strips), parked vehicles on the roadside, traffic flow, and bad road surfaces (rainy muddy / snow / construction sections).

  That is, for example, among the grouped data point sequences arranged in the traveling direction of the vehicle, the length is about 10 m or more and the angle change between the end points of the point sequence is 45 ° or less and the road side (left side) is protected. Identify the fence, and identify the center of the road (right side) as the median strip.

  As another example of the guard fence judgment method, for example, as shown in FIG. 4, the guard fence is composed of struts arranged at equal intervals and pipes connecting them, so that the reflection intensity of the point sequence is large. It is also possible to specify a guard fence that has a length that changes periodically (approximately every 2 m) according to the interval between the columns, or a point sequence that is 2 m apart in the road extension direction.

  Further, in recent years, there is a reflector provided with a reflective reflector at the upper part of the support fence, and the reflective reflector increases the reflection of laser light. For this reason, it is also possible to identify the protective fence by comparing the intensity of the reflected laser light with a predetermined threshold value to determine the reflective reflector.

  Further, the barrier detection unit 46 detects a point sequence of 1.5 m or more in length from a candidate of each object obtained by removing the guard fence and the central separation band from each object grouped by the grouping processing unit 40 (two-wheeled vehicle). And the stationary vehicle on the road side is identified as a parked vehicle.

  In addition, when a preceding vehicle with a short inter-vehicle distance or a vehicle on the opposite lane is moving at a certain speed or more, it is regarded as a barrier caused by traffic flow that obstructs the crossing of pedestrians.

  In addition, barriers such as guard fences and median strips are fixed at a fixed position for a long period of time. Therefore, the position of the barrier is registered in the map information stored in the HDD 28, and GPS (Global Positioning System) or the like is registered. You may estimate the position of the said barrier from the own vehicle position obtained by the positioning means.

  The guard fence detected by scanning with the laser radar 12 or imaging with the camera may be registered in the record type map information, and when traveling again in the same place, the map information may be read out and used for barrier determination. .

  Furthermore, the barrier detection part 46 is good also as what detects a bad road surface as a barrier. For example, on a rough road surface due to snowfall or rain, the reflection of laser light is larger than in normal times, and the dispersion of distances of distance measurement results by stereo viewing of images obtained by scanning with the laser radar 12 or imaging with a camera increases. Therefore, it is possible to detect a rough road surface by detecting the dispersion of the distance of the distance measurement result. In addition, the color and texture of road surfaces that are in rough road conditions such as the color and texture of gravel roads and gravel are stored in advance, and the images obtained by imaging with the camera and the color of each road surface stored in advance.・ Similarity is derived by comparing with each texture, and the color with the highest similarity ・ The road condition corresponding to the texture may be detected as the condition of the road surface. -It is good also as what determines with a rough road surface when similarity differs from a texture by a predetermined amount or more.

  The collision risk estimation unit 48 collides with the vehicle when the detected pedestrian enters the vehicle path from the crossing intention estimated by the crossing intention estimation unit 44 and the barrier position detected by the barrier detection unit 46. The possibility of collision is calculated, the collision risk with the host vehicle is calculated, and a warning is issued to the user from the display 14 and the speaker 34 according to the calculated risk.

  By the way, the process by each component (Grouping process part 40, the pedestrian detection part 42, the crossing intention estimation part 44, the barrier detection part 46, and the collision risk estimation part 48) of the driver assistance apparatus 10 comprised as mentioned above. Can be realized by a software configuration using a computer by executing a program. However, the present invention is not limited to realization by a software configuration, and needless to say, it can also be realized by a hardware configuration or a combination of a hardware configuration and a software configuration.

  Below, the case where the driver assistance apparatus 10 which concerns on this Embodiment implement | achieves the process by said each component by executing a collision risk estimation process program is demonstrated. In this case, the collision risk estimation processing program is installed in advance in the ROM 26 or HDD 28, provided in a state stored in a computer-readable recording medium, or distributed via wired or wireless communication means. The form etc. which are made can be applied.

  Next, with reference to FIG. 5, the operation of the driver support apparatus 10 according to the present embodiment will be described. FIG. 5 is a flowchart showing a flow of processing of the collision risk degree estimation processing program executed by the CPU 22. The collision risk degree estimation processing program is executed by the CPU 22 when a processing start instruction signal is received from a vehicle control device (not shown), for example, when an engine switch of a vehicle on which the driver assistance device 10 is mounted is turned on. .

  In step 100 of FIG. 8, the input of situation information is waited. In the next step 102, the position of the three-dimensional object indicated by the inputted situation information is plotted, and the region where the three-dimensional object continues is grouped as an object. To do.

  In the next step 104, the three-dimensional object of the predetermined size about the width of the pedestrian is tracked in the time direction as the pedestrian candidate from each object grouped in the step 102, and the predetermined amount of the pedestrian is moved. A person moving on the road at a speed is identified as a pedestrian, and the position of the pedestrian relative to the host vehicle, the moving speed of the pedestrian, and the moving direction of the pedestrian are detected. In step 104, pattern identification is performed on the area detected as a pedestrian to detect the pedestrian's posture.

  In the next step 106, the position and posture of the pedestrian detected in the above step 104, the surrounding bus stop or station, school, store, company location indicated by the map information stored in the HDD 28, the current time, etc. Estimate the pedestrian's intention to cross by finding the sum of contributions for each item. In this step 106, as the pedestrian's intention to cross, the crossing probability pc that the pedestrian crosses in the crossing direction with respect to the road on which the vehicle travels, and the translation probability pt that the pedestrian moves in the translational direction with respect to the road. This is derived as pc + pt = 1.

  In the next step 108, a barrier that prevents the movement of a pedestrian such as a structure on the road, a parked vehicle on the road, a traffic flow, and a bad road surface is detected from each object grouped in step 102.

  When such a barrier exists between the roadway and the sidewalk (roadside), the pedestrian is prevented from entering the vehicle lane. For this reason, the possibility of sudden jump-out of pedestrians and bicycles is reduced by the barrier. A typical barrier is a roadside guard fence.

  On the other hand, when a barrier such as a roadside parked vehicle or a bad road surface is present on the sidewalk or the roadside, the pedestrian is prevented from translating along the road. For this reason, the possibility that a pedestrian who is translating along the road protrudes into the roadway and enters the roadway increases.

  In the next step 110, the possibility of the detected pedestrian entering the vehicle path and colliding with the vehicle is predicted from the crossing intention estimated in step 106 and the position of the barrier detected in step 108. Then, the collision risk with the own vehicle is calculated.

Here, in the collision risk estimation process, the collision probability for the pedestrian is assumed as follows.
-If there is a barrier that prevents crossing between the pedestrian and the roadway, the crossing probability pc that the pedestrian crosses the roadway is assumed to be minimal. In the following calculation, the crossing probability pc is set to 0 for convenience of calculation.
・ If there is no barrier between the pedestrian and the vehicle lane, it is assumed that the pedestrian crosses with the estimated probability.
・ If there is a barrier in the direction in which the pedestrian translates on the road, it is assumed that the pedestrian protrudes along the barrier edge.

  Based on the above assumption, the collision probability that a physical collision occurs is calculated. At this time, the collision probability between the pedestrian jumping ahead of the vehicle and the vehicle depends on the relative position and relative speed between the vehicle and the pedestrian.

  Here, the variables are defined as follows, where the crossing direction is the x direction and the traveling direction of the vehicle is the y direction with respect to the road on which the vehicle is traveling.

vc: Speed of own vehicle x0: Relative position with respect to own vehicle when predicting collision risk (crossing direction)
z0: Relative position with respect to own vehicle when predicting collision risk (traveling direction)
vp: Pedestrian movement speed, generally 1.2 m / s and setting td: Pedestrian crossing start delay time tc: Time when the vehicle approaches the pedestrian closest. That is, the time of passing or colliding with the side of the pedestrian xp: relative position of the vehicle and the pedestrian at the time tc (transverse direction)
w: width of own vehicle pc: probability of pedestrian crossing Δvp: standard deviation of pedestrian crossing speed Δtd: standard deviation of delay time of pedestrian crossing start

  Here, for example, when the pedestrian is moving in the transverse direction with respect to the road and the vehicle and the pedestrian at the time of collision risk prediction are in a positional relationship as shown in FIG. The positional relationship between the vehicle and the pedestrian (see FIG. 6B) can be calculated by the following equation (1).

  xp = x0 + vp × (z0 / vc−td) Expression (1)

  At this time tc, a collision between the vehicle and the pedestrian occurs when the front surface of the vehicle and the pedestrian overlap, that is, when -w / 2 <xp <w / 2 is satisfied. Here, it is assumed that the size of the pedestrian is negligible compared to the vehicle. When calculating with a margin, a set value (pedestrian width equivalent amount 0.4 m to a safe distance margin between the pedestrian and the vehicle is about 1 m) is added to w.

  Further, the standard deviation σ (xp) of the pedestrian position xp at time tc can be calculated by the following equation (2).

  Therefore, when the predicted position of the pedestrian at time tc is calculated by the above formula (1) and the standard deviation is calculated by the above formula (2), the collision probability p is expressed by the area shown in FIG.

When the probability density function of the normal distribution is expressed by N (μ, σ ² ) as an average μ and variance σ ² , the collision probability p at this time is expressed by the following equation (3).

  On the other hand, for example, the pedestrian is moving in the translational direction (traveling direction) with respect to the road, and the positional relationship between the vehicle and the pedestrian at the time of collision risk prediction is as shown in FIG. In some cases, the presence of a barrier (roadside parked vehicle / bad road surface) that prevents movement in the translational direction causes the pedestrian to protrude from the sidewalk to the roadway, and as shown in FIG. The collision probability p with the pedestrian is calculated on the assumption that the vehicle moves along the side edge of the roadway.

  When the translational movement of the pedestrian is assumed, the time tc when the host vehicle passes the position of the pedestrian and the zp at the time of passing are obtained.

  It is assumed that the pedestrian in the translation direction is already walking at the moving speed vp, and the moving speed is known by observation by sensing or the like.

  At this time, the relationship of the following formulas (4) and (5) is satisfied.

  tc = zp / vc Expression (4)

  zp = z0 × vc / (vc−vp) (5)

  The position of the roadway side edge of the barrier at zp is obtained from the barrier detection result, and the collision probability is calculated by the above equation (3) with the position as the lateral position xp of the pedestrian at time tc. At this time, σ (xp) is the standard deviation of the amount of wobbling in the lateral direction of the pedestrian, and is set from the observation result of the pedestrian.

  And the final collision probability P is calculated | required from following Formula (6).

  P = pc × pcc + pt × ptc (6)

  This pc is the pedestrian's crossing probability, pt is the pedestrian's translation probability, pcc is the crossing direction collision probability p between the pedestrian and the own vehicle, and ptc is the translation direction of the pedestrian and the own vehicle. Is the collision probability p.

  In the next step 112, it is determined whether or not the collision probability P estimated in step 110 is larger than a predetermined reference value. If the determination is affirmative, the process proceeds to step 114, while the negative determination is made. If so, the process proceeds to step 116.

  In step 114, a warning sound is generated from the speaker 34, and an appropriate vehicle speed is displayed on the display 14. As for the appropriate vehicle speed, the collision probability P is recalculated assuming a plurality of vehicle speeds vc, and the maximum value of the vehicle speed that is equal to or less than the reference value is set.

  Considering the resolution of the vehicle speed presented to the driver, the vehicle speed on a general road is at most several stages, such as 60 km / h, 40 km / h, and 20 km / h.

  It should be noted that the appropriate vehicle speed may be set so that the value corrected by adding the weight to the collision probability p is kept below the reference value. The weight of the pedestrian can be considered as this weight. In particular, in a pedestrian accident, the degree of obstacle that a pedestrian receives depends on the vehicle speed vc. Therefore, simply, the term of the square of the vehicle speed proportional to the kinetic energy of the vehicle may be added to the collision probability.

  Therefore, for example, the weighted collision probability D can be calculated by the following equation (7).

  At this time, vs is a reference speed, and is set to 60 km / h for a domestic general road.

  Further, the weight based on the damage level may be set by statistically investigating the relationship between the vehicle speed and the damage level at the time of collision from actual accident statistical data.

  In the next step 116, for example, it is determined whether or not a processing end instruction signal is input from a vehicle control device (not shown), for example, when the vehicle engine switch is turned off. If the determination is negative, the process proceeds to step 100. On the other hand, if the determination is affirmative, the process of the collision risk degree estimation processing program ends.

  As described above, according to the present embodiment, the position of the pedestrian moving around the traveling vehicle, the moving speed of the pedestrian, and the moving direction of the pedestrian are detected, and the movement of the pedestrian is prevented. Collision risk that pedestrian and vehicle collide based on detected pedestrian position, moving speed, and moving direction by detecting the position of barrier and taking into account the detected barrier position for pedestrian and vehicle Since the degree is estimated, the collision risk can be estimated with high reliability.

  In the present embodiment, the case of presenting the appropriate vehicle speed to the driver by displaying the appropriate vehicle speed on the display 14 has been described. However, the present invention is not limited to this, and for example, as follows: Can be considered.

・ Display the appropriate vehicle speed on the meter.
・ Encourage the driver to slow down by voice.
・ Alarm if the current vehicle speed is higher than the appropriate vehicle speed.
・ If the current vehicle speed is higher than the appropriate vehicle speed, the vehicle brake is controlled to perform intervention braking to the appropriate vehicle speed.

  Moreover, in this Embodiment, based on the estimated collision probability P, you may make it implement the pedestrian warning only for the pedestrian with the said collision probability P larger than the said reference value. Thereby, the excessive warning with respect to the pedestrian who exists in the road periphery can be suppressed, and a pedestrian warning can be implemented appropriately.

  Similarly, when the estimated collision risk P is larger than the reference value, control may be performed to increase the braking amount of the brake with respect to the brake operation to the brake pedal provided in the vehicle. . Thereby, excessive operation of the pre-crash safety (PCS) device can be reduced, and appropriate brake assist can be realized.

  In addition, when calculating with the above formula (3), recalculate the risk when x0 is changed to ± 0.5 m and ± 1 m, respectively, and if the risk is low, You may make it show to a driver so that the distance of the horizontal direction of the own vehicle may be opened.

  In addition, the barrier detection unit 46 outputs information on the road area where the pedestrian enters the barrier at a low frequency to the pedestrian detection unit 42 based on the detection result of the barrier, and the pedestrian detection unit 42 The road area where the intrusion frequency is low may be masked from the detection area of the pedestrian. Thereby, since the detection area | region of a pedestrian can be reduced, the erroneous detection and calculation cost of a pedestrian can be reduced.

  In addition, the main configuration (see FIG. 1) of the electrical system of the driver support apparatus 10 described in the present embodiment and the functional configuration (see FIG. 3) of the driver support apparatus 10 are examples, and the present invention. Needless to say, changes can be made as appropriate without departing from the spirit of the present invention.

  Further, the flow of processing of the collision risk estimation processing program (see FIG. 5) described in the present embodiment is also an example, and it goes without saying that it can be changed as appropriate without departing from the gist of the present invention.

It is a block diagram which shows the principal part structure of the electric system of the driver assistance apparatus which concerns on embodiment. It is a figure which shows an example of the detection result of the surrounding condition by the laser radar which concerns on embodiment. It is a block diagram which shows the functional structure of the driver assistance apparatus which concerns on embodiment. It is the figure which showed typically the detection of the protection fence by the laser radar which concerns on embodiment. It is a flowchart which shows the flow of a process of the collision risk estimation process program which concerns on embodiment. It is a figure which shows an example of the positional relationship of the pedestrian who moves to the crossing direction which concerns on embodiment, and a vehicle. It is a figure which shows an example of the standard deviation of the position of the pedestrian which concerns on embodiment. It is a figure which shows an example of the positional relationship of the pedestrian who moves to the translation direction which concerns on embodiment, and a vehicle.

Explanation of symbols

10 Driver support device 12 Laser radar (acquisition means)
14 Display (display means)
34 Speaker (Warning means)
42 Pedestrian detection unit (pedestrian detection means)
44 Crossing intention estimation unit (estimation means, position information acquisition, derivation means)
46 Barrier detector (barrier detector)
48 Collision risk estimation part (estimation means)

Claims (13)

Pedestrian detection means for detecting the position of the pedestrian moving around the traveling vehicle, the moving speed of the pedestrian and the moving direction of the pedestrian,
Barrier detection means for detecting the position of the barrier that prevents movement of the pedestrian;
The pedestrian and the vehicle based on the position, moving speed, and moving direction of the pedestrian detected by the pedestrian detection means, and the position of the barrier detected by the barrier detection means for the pedestrian and the vehicle. An estimation means for estimating a collision risk of collision with
A collision risk estimation device comprising: The image processing apparatus further comprises acquisition means for acquiring situation information indicating the surrounding situation of the vehicle by performing at least one of imaging around the vehicle, scanning with laser light, and scanning with radio waves,
The pedestrian detection means detects the position of the pedestrian, the movement speed of the pedestrian and the movement direction of the pedestrian from the surrounding situation indicated by the situation information acquired by the acquisition means,
The collision risk estimation device according to claim 1, wherein the barrier detection unit detects the position of the barrier from the surrounding situation indicated by the situation information acquired by the acquisition unit. The collision risk estimation device according to claim 1, wherein the barrier is at least one of a protective fence, a structure including a median, and a structure that prevents movement of the pedestrian, a traffic state of an oncoming lane, and a rough road surface. . When the barrier detected by the barrier detection means is located between the pedestrian and the vehicle, the estimation means has a higher collision risk than when the barrier is not located between the pedestrian and the vehicle. The collision risk estimation device according to any one of claims 1 to 3, wherein the collision risk estimation device is estimated to be low. The estimating means is such that the moving direction of the pedestrian detected by the pedestrian detecting means is substantially parallel to the moving direction of the vehicle, and the barrier detected by the barrier detecting means is in front of the moving direction of the pedestrian. The collision risk estimation device according to any one of claims 1 to 4, wherein the collision risk is estimated to be higher when the barrier is positioned than when the barrier is not positioned forward in the movement direction. The pedestrian detection means further detects the body direction or the face direction of the pedestrian,
In the case where the direction of the body or face of the pedestrian detected by the pedestrian detection unit is the front or the back of the vehicle, the direction of the estimation is the front or the back of the vehicle. The collision risk estimation device according to any one of claims 1 to 5, wherein the collision risk is estimated to be lower than in a case other than the above. It further comprises position information acquisition for acquiring position information indicating a position of a predetermined use target that can be used by a pedestrian,
When the position of the utilization target indicated by the position information acquired by the position information acquisition is located on the opposite side across the road on which the vehicle travels, the estimation means is The collision risk estimation apparatus according to any one of claims 1 to 6, wherein the collision risk is estimated to be higher than when the position of the use target is not located on the opposite side. The said estimation means presumes a collision risk higher when it is a time zone where the said utilization object is used frequently by the said pedestrian than the case other than the said time zone used frequently. Collision risk estimation device. The estimation means calculates a cross-direction collision probability that the pedestrian moves in a crossing direction with respect to a road on which the vehicle travels and the pedestrian and the vehicle collide, and a translational direction with respect to the road A translation direction collision probability that the pedestrian moves and the pedestrian collides with the vehicle is calculated, and the collision risk is estimated by adding the transverse direction collision probability and the translation direction collision probability. The collision risk estimation device according to any one of claims 1 to 8. The said pedestrian detection means excludes the area | region where the movement of the said pedestrian is prevented by the said barrier detected by the said barrier detection means from the detection object of a pedestrian. Collision risk estimation device. The collision risk estimation device according to any one of claims 1 to 10,
Deriving means for deriving the speed of the vehicle at which the collision risk estimated by the collision risk estimation device is a predetermined value or less as an appropriate speed;
Display means for displaying the appropriate speed derived by the deriving means;
A driver support device comprising: The collision risk estimation device according to any one of claims 1 to 10,
Warning means for warning the driver of the vehicle when the collision risk estimated by the collision risk estimation device is larger than a predetermined value;
A driver support device comprising: The collision risk estimation device according to any one of claims 1 to 10,
Braking for controlling to increase the braking amount of the brake with respect to the brake operation to the brake pedal provided in the vehicle when the collision risk estimated by the collision risk estimation device is larger than a predetermined value. A quantity control means;
A driver support device comprising:

Images